U.S. patent number 7,933,679 [Application Number 12/256,257] was granted by the patent office on 2011-04-26 for method for analyzing and optimizing a machining process.
This patent grant is currently assigned to Cessna Aircraft Company. Invention is credited to Amit Deshpande, Dhananjay Joshi, Parag Konde, Pravin Kulkarni, Kareem Syed.
United States Patent |
7,933,679 |
Kulkarni , et al. |
April 26, 2011 |
Method for analyzing and optimizing a machining process
Abstract
A method for optimizing machining parameters for a cutting
process performed on a work piece. Finite element analysis of
cutting tool and work material interaction is initially performed.
Mechanistic modeling of the cutting process, using results of the
finite element analysis, is then performed to provide optimized
machining parameters for improved rate of material removal and tool
life. Optionally, a two-stage artificial neural network may be
supplementally employed, wherein a first stage of the network
provides output parameters including peak tool temperature and
cutting forces in X and Y directions, for a combination of input
reference parameters including tool rake angle, material cutting
speed, and feed rate.
Inventors: |
Kulkarni; Pravin (Wichita,
KS), Joshi; Dhananjay (Wichita, KS), Konde; Parag
(Wichita, KS), Deshpande; Amit (Wichita, KS), Syed;
Kareem (Wichita, KS) |
Assignee: |
Cessna Aircraft Company
(Wichita, KS)
|
Family
ID: |
43880508 |
Appl.
No.: |
12/256,257 |
Filed: |
October 22, 2008 |
Related U.S. Patent Documents
|
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
Issue Date |
|
|
60982027 |
Oct 23, 2007 |
|
|
|
|
Current U.S.
Class: |
700/173; 700/180;
703/2 |
Current CPC
Class: |
G05B
13/0265 (20130101); G05B 2219/36252 (20130101) |
Current International
Class: |
G06F
19/00 (20110101) |
Field of
Search: |
;700/175,172,182,180,177,173 ;703/2,56 ;340/683 ;73/579 |
References Cited
[Referenced By]
U.S. Patent Documents
Other References
Mackerle,Jaroslav, Finite-element analysis and simulation of
machining, Jun. 26, 1999, Elsevier Sciences, p. 17-23. cited by
examiner .
Albert J. Shih, Finite Element Analysis of the Rake Angle Effects
in Orthogonal Metal Cutting, Jan. 27, 1995, Elsevier Science p.
1-17. cited by examiner .
Karpat, Yigit, et al. "Multi-Objective Optimization for Turning
Processes Using Neural Network Modeling and Dynamic-Neighborhood
Particle Swarm Optimization," Int. J. Adv. Manuf Technology, DOI
10.107/s00170-006-0719-8, 14 pages, Feb. 24, 2006. cited by other
.
Zuperl, Uros, et al. "Optimization of Cutting Conditions During
Cutting by Using Neural Networks," Robotics and Compuer Integrated
Manufacturing 19 (2003), pp. 189-199. cited by other.
|
Primary Examiner: Bahta; Kidest
Attorney, Agent or Firm: Lathrop & Gage LLP
Parent Case Text
RELATED APPLICATION
This application claim priority to provisional Patent Application
Ser. No. 60/982,027, filed Oct. 23, 2008, the disclosure of which
is incorporated herein by reference.
Claims
What is claimed is:
1. A method for optimizing machining parameters for a milling
process performed on a work piece using a cutting tool, the method
comprising: performing finite element analysis of cutting tool and
work material interaction to determine an optimized combination of
the machining parameters by varying work material properties,
cutting tool geometry, cutting speed and feed rate; performing
mechanistic modeling of the milling process, using results of the
finite element analysis, to provide optimized machining parameters
for improved rate of material removal and tool life; using a
two-stage artificial neural network, wherein a first stage of the
network provides output parameters including peak tool temperature
and cutting forces in X and Y directions, for a combination of
input reference parameters including tool rake angle, material
cutting speed, and feed rate; wherein the output parameters
obtained from the first stage are considered to be baseline
parameters; applying acceptance criteria to the baseline parameters
determine acceptable baseline parameters; using the acceptable
baseline parameters as input to a second stage of the network;
obtaining optimized values of machining parameters, including
surface speed and feed rate, from the second stage of the network;
and utilizing the optimized values to improve work piece material
removal rate and cutting tool life.
2. The method of claim 1, wherein resonance-based vibration
analysis is performed prior to the finite element analysis and the
mechanistic modeling, by analyzing cutting tool operation to
determine machining parameters which result in reduced tool
chatter.
3. The method of claim 2, wherein the optimized combination of
machining parameters include parameters selected from the set
consisting of work piece material grade, set-up rigidity, cutting
tool material, flute type, insert type, corner radius, and number
of flutes.
4. The method of claim 1, wherein: the finite element analysis is
performed by simulating operation of the cutting tool to determine
the optimized cutting speed and feed rate; and the mechanistic
modeling is performed by analyzing operation of the cutting tool to
balance resultant cutting forces, temperature, and milling machine
power.
5. The method of claim 1, wherein the finite element analysis is
performed using information including parameters selected from the
set consisting of machining parameters, work material properties
and cutting tool details, to provide a prediction of parameters
selected from the group consisting of cutting forces, tool
temperature, tool stresses, tool pressure, and milling machine
power consumption, resulting from machining the work piece.
6. The method of claim 5, wherein the finite element analysis is
performed by varying cutting speed and feed rate while maintaining
constant depth of cut.
7. The method of claim 1 wherein the mechanistic modeling is
performed using parameters selected from the set consisting of work
piece material properties, raw stock geometry, cutting tool
details, and numerical control tool path to provide a prediction of
cutting force, tool temperature, and milling machine power
consumption, resulting from machining various features of the work
piece.
8. A method for optimizing machining parameters for a turning
process performed on a work piece using a cutting tool, the method
comprising: performing finite element analysis of cutting tool and
work material interaction to determine an optimized combination of
the machining parameters, using a finite element model, by varying
work material properties, cutting tool geometry, cutting speed and
feed per tooth; performing mechanistic modeling of the turning
process, using results of the finite element analysis, to provide
optimized machining parameters for improved rate of material
removal and tool life; using a two-stage artificial neural network,
wherein a first stage of the network provides output parameters
including peak tool temperature and cutting forces in X and Y
directions, for a combination of input reference parameters
including tool rake angle, material cutting speed, and feed per
tooth; wherein the output parameters obtained from the first stage
are considered to be baseline parameters; applying acceptance
criteria to the baseline parameters determine acceptable baseline
parameters; using the acceptable baseline parameters as input to a
second stage of the network; obtaining optimized values of
machining parameters, including surface speed and feed per tooth,
from the second stage of the network; and utilizing the optimized
values to improve work piece material removal rate and cutting tool
life.
9. The method of claim 8, wherein the parameters to be optimized
are selected from the set consisting of surface speed, feed rate,
depth of cut, tool material, holder types, and insert types.
10. The method of claim 8, wherein: the finite element analysis is
performed by simulating operation of the cutting tool to determine
the optimized cutting speed and feed rate; and the mechanistic
modeling is performed by analyzing operation of the cutting tool to
balance resultant cutting forces, temperature, and turning machine
power.
11. The method of claim 8, wherein the finite element analysis is
performed using information including parameters selected from the
set consisting of machining parameters, work material properties
and cutting tool details, to provide a prediction of parameters
selected from the group consisting of cutting forces, tool
temperature, tool stresses, tool pressure, and turning machine
power consumption, resulting from machining the work piece.
12. The method of claim 11, wherein the finite element analysis is
performed by varying cutting speed and feed rate while maintaining
constant depth of cut.
13. The method of claim 8 wherein the mechanistic modeling is
performed using parameters selected from the set consisting of work
piece material properties, raw stock geometry, cutting tool
details, and numerical control tool path to provide a prediction of
cutting force, tool temperature, and turning machine power
consumption, resulting from machining various features of the work
piece.
14. A method for optimizing machining parameters for a cutting
process performed on a work piece using a cutting tool, the method
comprising: performing finite element analysis of cutting tool and
work material interaction; performing mechanistic modeling and
analysis of the cutting process, using results of the finite
element analysis, to provide machining parameters for an improved
rate of material removal and tool life; determining an optimized
combination of the machining parameters, using a finite element
model, by varying work material properties, cutting tool geometry,
cutting speed and feed rate; using a two-stage artificial neural
network, wherein a first stage of the network provides output
parameters including peak tool temperature and cutting forces in X
and Y directions, for a combination of input reference parameters
including tool rake angle, material cutting speed, and feed rate;
wherein the output parameters obtained from the first stage are
considered to be baseline parameters; applying acceptance criteria
to the baseline parameters determine acceptable baseline
parameters; using the acceptable baseline parameters as input to a
second stage of the network; obtaining optimized values of
machining parameters, including surface speed and feed rate, from
the second stage of the network; and utilizing the optimized values
to improve work piece material removal rate and cutting tool life.
Description
TECHNICAL FIELD
This application relates to analysis of machining processes,
including milling, drilling, turning and boring, using computer
simulation and analysis techniques. More particularly, this
invention relates to optimization of the above-mentioned machining
processes by combining and analyzing results from finite element
analysis, mechanistic modeling and vibration analysis of those
processes.
PROBLEM AND SOLUTION
The traditional approach followed in selection of machining process
parameters is a trial and error method based on machining handbooks
and machinist's experience. This traditional approach of parameter
selection is not specific to a particular machine tool and
typically does not provide optimum utilization of cutting tool and
machine tool capacity, which extends product manufacturing cycle
time significantly. The present method involves machining process
optimization using finite element analysis, mechanistic modeling
and optional vibration analysis to overcome the disadvantages of
the traditional approach. The present approach is economical, as
the optimization process is carried out via computer simulation
instead of in a machine shop, which eliminates need for time
consuming shop trials and the cost of cutting tools and material.
The data generated through the computer simulation is then
validated in the machine shop and is documented in a structured
database.
Machining process analyses are carried out using commercially
available finite element analysis software, mechanistic process
modeling software, and vibration analysis software and hardware in
combination. The analyses reduce machining time, increase cutting
tool life, improve productivity, improve part quality and optimize
utilization of cutting tool and machine tool capacity. The present
process combines the results of these analyses in order to
determine optimum machining parameters. The output parameters of
the simulation techniques include cutting forces, peak tool
temperature and power consumption. Machining process parameters to
be optimized using this methodology include cutting speed, feed
rate, depth of cut and radial engagement of cutter. These analysis
techniques result in optimization of the complete machining system,
which includes cutting tool, work piece material, and machine tool.
These simulation techniques and their combined application increase
the understanding of the mechanics of machining processes and
reduce the number of trial and error experiments.
SUMMARY
The present method provides optimized machining parameters for a
cutting process performed on a work piece. Finite element analysis
of cutting tool and work material interaction is initially
performed. Mechanistic modeling of the cutting process, using
results of the finite element analysis, is then performed to
provide improved machining parameters for rate of material removal
and tool life.
In an alternative embodiment, a two-stage artificial neural network
may be supplementally employed, wherein a first stage of the
network provides output parameters including peak tool temperature
and cutting forces in X and Y directions, for a combination of
input reference parameters including tool rake angle, material
cutting speed, and feed rate.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a flowchart showing an exemplary high level view of the
present system and process;
FIG. 2A is a flowchart showing an exemplary set of steps performed
in analysis and optimization of a milling operation;
FIG. 2B is a diagram of a exemplary predictive configuration for
performing vibration analysis;
FIG. 3 is a flowchart showing an exemplary set of steps performed
in analysis and optimization of a drilling operation;
FIG. 4 is a flowchart showing an exemplary set of steps performed
in analysis and optimization of a turning/boring operation;
FIG. 5 is a diagram of a exemplary first-stage Artificial Neural
Network model for predicting the results of a machining operation;
and
FIG. 6 is a diagram of a exemplary first-stage Artificial Neural
Network model for predicting the results of a machining
operation.
DETAILED DESCRIPTION
The present system applies machining analysis techniques including
finite element analysis, mechanistic modeling, and vibration
analysis, individually or in combination, for analysis and
optimization of various machining processes. FIG. 1 is a flowchart
showing an exemplary high level view of the present system and
process 100. As shown in FIG. 1, machining processes analyzed and
optimized by the present system and method include milling,
drilling, turning, and boring. The techniques employed by the
present method assist in resolving machining issues such as low
production rate, underutilized machine capacity, low tool life,
tool vibrations or chatter, etc.
Analysis of a particular milling operation includes vibration
analysis, described in detail below, as indicated by step 105. This
analysis is performed using a set of software programs running on a
computer system 120 to troubleshoot tool chatter issues and
determine chatter-free machining parameters. Objectives of
vibration analysis include an improved surface finish on a
particular part, balancing of the forces on the machine tool, and
enhancing the life of cutting tool. After vibration analysis is
performed for the milling operation, process flow continues with
finite element method analysis and mechanistic modeling, as
described below.
Analysis of a drilling and turning/boring operations begins with
finite element method (FEM) analysis. Parameters affecting
machining processes include cutting forces and associated stresses
on the cutting tools and in the machined work piece, temperature
distribution in the tool and on the tool surfaces and the chip
formation. Each of these parameters may be analyzed using finite
element simulations to determine optimum cutting speed and feed
rate, at step 110.
Compared to empirical and analytical methods, finite element method
analysis has advantages in several respects. FEM analysis processes
material properties as a function of strain, strain rate and
temperature. Commercially available finite element analysis
software is used to simulate machining processes. The analysis is
conducted to evaluate the effect of changes in process parameters
on the process outputs including cutting speed and cutting forces,
peak tool temperature and estimated machine power requirement.
Finite element software analyzes the process at a micro level and
helps to optimize machining parameters and cutting tool life.
At step 115, commercially available mechanistic process modeling
software is used to analyze the process at a macro level. The
complete numerical control (NC) tool path is analyzed to balance
the resulting cutting forces, temperature and power required to
machine a particular part. Optimized machining parameters derived
from these techniques are applied in a numerical control program
and are documented in a database 150 (connected to computer 120)
after successful implementation.
Machining database 150 is a knowledge base that is used to store
the information on the tools, work piece material, set-up,
operation, conventional and optimized machining parameters, and
finite element analysis results from the simulation. Users can
look-up optimized machining parameters for the machining conditions
they are encountering and troubleshoot problems that may arise
during the machining process. Data entry includes tool information,
work material, machine information, machining parameters, and
finite element analysis (FEM) results.
The tool information input to machining database 150 includes
information related to tool holder assembly identification,
manufacturer information, tool details such as material and
geometry. The work information includes work material and grade,
hardness, and raw stock type. The operation information includes
set-up rigidity grading, operation type, type of cut. Set-up
information includes machine number the tool is used on and whether
it is a standard tool on the machine or not. Input data includes
minimum and maximum conventional machining parameters as well as
optimum machining parameters. Finite element analysis results
include cutting force, thrust force, radial force, temperature,
expected tool life, material/operation/machine information,
machining parameters, and FEM results.
Optimization Methodology for Milling Operations
FIG. 2A is a flowchart showing an exemplary set of steps performed
in analysis and optimization 200 of a milling operation. As shown
in FIG. 2A, the present process for optimizing milling operations
includes analysis of machining parameters, tool life, tool
selection, and tool failure investigation. Information about
milling machine spindle speed (RPM), feed rate, radial width of
cut, axial depth of cut, work piece material grade, set-up
rigidity, and cutting tool material, flute type, insert type,
corner radius, and number of flutes, is initially collected at step
203. A selected tool is then checked for possibility of chatter or
vibration during the machining operation at step 205. The criteria
for the tool chatter test includes the predetermined value of tool
length to diameter ratio, part feature characteristics, and fixture
conditions. If the tool is expected to chatter, then it is analyzed
to determine the chatter-free machining parameters using the
frequency response of the cutting tool, spindle, and machine
system.
Vibration analysis may be used to optimize and troubleshoot milling
processes. A commercially available vibration analysis package,
such as MetalMax, distributed by Technicut Ltd., includes vibration
analysis software. The present system uses the predictive module of
the analysis package. Using the predictive module, a static
frequency response test is performed on the milling cutter by
placing it in the stationary spindle of the machine.
FIG. 2B is a diagram of a exemplary configuration 201 for
performing vibration analysis. As shown in FIG. 2B, analysis is
conducted by a `tap test` in which the milling cutter 204 is tapped
using a test hammer 202 and the frequency response of the system is
received through an accelerometer 206. The hammer 202 and
accelerometer 206 are connected to a data acquisition/control unit
208.
As indicated in FIG. 2A, at step 210, a frequency lobe diagram is
generated in accordance with the procedure described above, which
allows determination of stable and unstable machine spindle speeds
and depth of cuts. If tool chatter is not expected, then the tool
is analyzed using FEM software, to determine optimum cutting speed
and feed rate, in step 215.
Finite Element Analysis
To analyze milling operations, three-dimensional multipoint milling
operations are simulated as two-dimensional single point turning
operations to reduce the simulation time. Process parameters, such
as cutting speed and feed rate, are converted from a multipoint
cutting condition to a single point cutting condition.
Two-dimensional turning and boring operations are simulated without
any other changes in the actual process. Drilling operations are
analyzed using a three dimensional process model.
The present machining process optimization involves analysis of
baseline machining parameters for a tool/work-piece material
combination using finite element simulation software such as
AdvantEdge.TM., available from Third Wave Systems, Inc. The results
of the baseline simulation provide a reference for optimization of
the machining parameters. The baseline simulation is followed by a
set of simulations performed via FEM by varying cutting speed and
feed rate while maintaining constant depth of cut. The objective of
these simulations is to determine machining parameters which
provide high material removal rate while maintaining reasonable
tool life.
The results of all the simulations, including cutting forces, peak
tool temperature, tool stresses, tool pressure, and power
consumption, are documented in a standard template. The results of
baseline simulations are used to establish an acceptance limit for
resulting cutting forces and peak tool temperature of the FEM
simulations. The limits for peak tool temperature are also set by
selection of the tool material. Machining parameters having results
within acceptable limits are selected as optimized for machining.
These optimized parameters are then used in the NC (numerical
control) program which controls the milling machine and are
validated by shop trials in the machine shop.
The results of the FEM analyses are combined to obtain optimum
machining parameters in step 220. The original NC program is
updated in step 225 by incorporating the optimized machining
parameters obtained in step 220. The updated NC tool path is then
analyzed for non-productive machining time, cutting forces, and
temperature, using mechanistic modeling software, in step 230.
Mechanistic Modeling
A mechanistic modeling technique is used to analyze milling,
drilling, turning and boring operations independently or in
combination with FEM and vibration analysis. Commercially available
mechanistic modeling software, such as Production Module.TM.,
available from Third Wave Systems, Inc., can be used for
two-dimensional and three-dimensional simulations of the
above-described machining processes. In contrast to FEM,
mechanistic modeling provides a complete tool path can be analyzed
for a series of operations by a cutter or set of cutters in
sequential order. In this method, material properties are assigned
to raw stock of the exact same shape and size as actual raw stock
to be machined. A coordinate system is established to reflect the
machine coordinate system, and geometric orientation of the raw
stock is matched with that in the NC program, which is provided to
the process simulation software. Mechanistic modeling is performed
using information including work piece material properties, raw
stock geometry, cutting tool details, and numerical control tool
path. The shape of the final component is generated after
completion of all the simulation operations. Results of the
simulation include cutting forces, temperature, and machine power
consumption, during machining of various features of the machined
component or work piece. This data is analyzed to set an acceptance
limit for the cutting forces and temperature. Machining parameters
are altered to improve overall material removal rate within the
bounds of cutting forces and temperature. The mechanistic modeling
software identifies non-productive passes or motions of the cutter,
which can be reduced or eliminated to improve overall machining
time. The selected machining parameters are listed in a
recommendation report. The machining process is analyzed and
optimized to reduce the machining cycle time and to utilize the
capacity of the cutting tool and machine tool to an optimum
extent.
The analysis results of the mechanistic modeling procedure are
compiled in step 235 to generate recommended optimized machining
parameters to update the NC program. The NC program is updated in
step 240 in accordance with the recommended parameters. After the
NC program is modified, it is validated in the machine shop to
evaluate the performance of the recommended tools using the
optimized machining parameters, in step 245. If problematic issues
related to tool life or part quality are experienced (step 247),
then feedback is provided for further analysis, and the present
process is repeated, continuing back at step 235. Finally, a
machining database (knowledge base) 150 is updated with the
optimized parameters for all the tools and corresponding machining
conditions at step 250.
Optimization Methodology for Drilling Operations
FIG. 3 is a flowchart showing an exemplary set of steps performed
in analysis and optimization 300 of a drilling operation. The
present process for optimizing drilling operations includes
analysis of machining parameters and tool life, tool selection, and
tool failure investigation. Information about cutting tools,
machine tool, work piece material grade and condition and set-up
rigidity is collected in step 303 for analysis. An initially
selected tool is analyzed using FEM software to determine optimum
cutting speed and feed rate in step 305. The NC program for the
drilling machine is updated in step 310 by incorporating the
optimum machining parameters obtained in step 305. The updated NC
tool path is then analyzed for the non-productive machining time,
cutting forces, and temperature using mechanistic modeling software
to determine optimized machining parameters, in step 315.
The NC program is updated with the optimized parameters in step
320. After the program is modified, it is validated in the machine
shop to evaluate the performance of the recommended tools with
optimized machining parameters, in step 325. If any problematic
issues related to tool life or part quality are experienced (step
327), then feedback is provided for further analysis, modified
parameters are determined, and the present process is repeated,
continuing back at step 310. Finally, the machining database 150 is
updated with the optimized parameters for the tools and
corresponding machining conditions at step 330.
Optimization Methodology for Turning and Boring Operations
FIG. 4 is a flowchart showing an exemplary set of steps performed
in analysis and optimization 400 of a turning/boring operation. The
present process for optimizing turning and boring operations
includes analysis of machining parameters and tool life, tool
selection, and tool failure investigation. Information about
cutting tools, machine tool, work piece material grade and
condition and set-up rigidity is collected in step 405 for
analysis. A selected tool is then analyzed using FEM software to
determine optimum cutting speed and feed rate in step 410. The NC
program for the turning/boring machine is updated in step 415 by
incorporating the optimum machining parameters obtained in step
410. The updated NC program is analyzed for the non-productive
machining time and cutting forces using the mechanistic modeling
software in step 420.
The NC program is updated in step 425 using the determined
optimized machining parameters including surface speed, feed per
revolution and depth of cut. After the NC program is modified, it
is validated in the machine shop to evaluate the performance of the
recommended tools at optimum machining parameters, in step 430. If
any problematic issues related to tool life or part quality are
experienced (step 433), then feedback is provided for further
analysis, modified parameters are determined, and the present
process is repeated, continuing back at step 415. Finally the
machining database 150 is updated with the optimized parameters for
the tools and corresponding machining conditions at step 435.
Machining Process Analysis Automation Using Artificial Neural
Networks
In an alternative embodiment, a machining process optimization
technique using finite element modeling (FEM) software may be used
to reduce machining time, to increase tool life, to improve
productivity and to optimize machine tool utilization. However,
this technique requires a significant amount of analysis time to
perform FEM simulations and to determine optimum machining
parameters.
Optimization of Machining parameters is a non-linear process with
constraints on allowable cutting temperature and cutting forces.
There are complex inter-relationships between machining parameters
and their collective effect on mechanics of machining. The use of
Artificial Neural Networks (ANNs) for machining process
optimization is one approach for automatic modeling of dynamic,
non-linear phenomena that are too complex to be solved by
conventional optimization algorithms. This methodology ensures
efficient and relatively fast selection of the optimum machining
parameters from an available set of pre-analyzed reference
parameters including work material properties, machining parameters
and cutting tool geometry.
Machining process optimization involves analysis of baseline
simulation of machining operations with conventional machining
parameters using a two-dimensional finite element method. The
baseline simulation is followed by FEM simulations in which cutting
speed and feed rate are varied to determine improved material
removal rate and reasonable cutting tool life. Depth of cut is kept
constant for all of the simulations. The results of all the
simulations are recorded in terms of cutting forces and peak tool
temperature. The results of baseline simulations are used to
establish acceptance limits for cutting forces and peak tool
temperature.
Machining parameters having results within the acceptable limits
are selected for recommendation. In order to shorten the analysis
time involved in this optimization procedure, automatic
optimization methodology of machining parameters using a two-stage
ANN for each cutting tool and work piece material combination is
employed in the present method.
Compared to conventional computing methods, neural networks are
more robust and generally better suited for analyzing complex sets
of information. Neural networks can be used to solve many nonlinear
optimization problems and work on the basis of feed-forward and
back-propagation algorithms which automatically adapt to current
conditions during training. The starting point for the use of a
neural network is representative data collected from finite element
simulation results through FEM simulations. In an exemplary
embodiment, ANN analysis is performed in steps 215 (FIG. 2A), 305
(FIG. 3), and 410 (FIG. 4) of the above-described embodiments.
Numerous FEM simulations are performed for different materials and
conditions by selecting values for inputs including tool geometry
and machining parameters within specified maximum and minimum
limits. Outputs of these simulations are used for training and
testing of Neural Networks.
Neural networks are trained using the simulation results of a
matrix of machining parameters. Network training involves the
process of interactively adjusting weights across the connections
of the network with the help of a back-propagation algorithm in
such a way that the prediction errors on the training set are
minimized.
Once the training is completed, testing of the network is done with
a set of simulation results (not used for network training) to
validate the outcome of neural network from supervised learning. If
the testing is successful and the error of prediction is within
permissible limits, then the empirical network model is fine-tuned
and made ready for use through a user interface.
Once the training and testing are successfully completed, the
empirical model 500 can be used to predict the results of
two-dimensional machining, namely, peak tool temperature, X-cutting
force and Y-cutting force for the given cutting speed, feed per
tooth and rake angle as input parameters.
FIG. 5 is a diagram of a exemplary first-stage artificial neural
network 500 for predicting the results of a machining operation
using finite element analysis of machining and tool geometry input
parameters. As shown in FIG. 5, information including tool rake
angle, material cutting (surface) speed, and feed per (cutting
tool) tooth is input to each of three neural network models 505,
506, and 507. These models respectively provide values for work
surface/tool temperature, X-plane cutting force, and Y-plane
cutting force.
Using the empirical model 500 shown in FIG. 5, FEM simulations are
performed for different materials and conditions by selecting
values for inputs including tool geometry (e.g., rake angle) and
machining parameters (e.g., cutting speed and feed rate) between
specified maximum and minimum limits.
FIG. 6 is an exemplary diagram showing an artificial neural network
system 600 used for continued finite element analysis of the
machining process and tool geometry input parameters. FIGS. 5 and 6
collectively comprise a two-stage artificial neural network
500/600. The results thus obtained from this model 500 are regarded
as baseline results and allowable limits (e.g., within 110% of
baseline results) are given as input parameters to a second
empirical model, which is the second-stage artificial neural
network 600 shown in FIG. 6, As shown in FIG. 6, output values from
artificial neural network 500 including tool rake angle,
temperature, and X- and Y-cutting force, are input to each of two
neural network models 605 and 606. Optimized values of surface
speed and feed per tooth are obtained as output parameters from the
second-stage ANN model 600 for improved material removal rate and
reliable tool life.
While preferred embodiments of the present method have been
described, so as to enable one of skill in the art to practice the
disclosed subject matter, the preceding description is intended to
be exemplary only, and should not be used to limit the scope of the
disclosure, which should be determined by reference to the
following claims.
* * * * *